Published: July 13, 2026

The defense workforce problem is not one problem, and treating it as one is how every tool fails at once.   Anduril, the upstart defense manufacturer best known for its autonomous drones, has published a manufacturing doctrine called Rebuild the Arsenal, and it may be the most consequential workforce document that nobody in workforce development is reading.  Stated plainly, the doctrine holds that America cannot deter a great-power war with small numbers of exquisite weapons that take a decade to build, and must instead mass-produce cheaper autonomous systems the way commercial industry produces cars and electronics, which is only possible if producibility is designed into the weapon from the first sketch.  The third of its design principles is the one that should stop a workforce reader cold, because it is workforce policy wearing an engineering badge.  Anduril wants weapons designed so they can be built by “the largest possible workforce” rather than by scarce specialists, which means stripping out the exotic materials, the artisanal processes, and the components only a cleared craftsman is allowed to touch, until the labor pool expands from the defense industrial base to the automotive plant and the electronics line.  The company means it.  The question this doctrine cannot answer, because it is not an engineering question, is what the rest of us are supposed to do about it.

What Anduril Actually Built

None of this is aspiration.  At the company’s Arsenal-1 plant in Ohio, production of the Air Force’s Fury combat drone already runs through 22 workstations of roughly equal workload, and Breaking Defense, which toured the line in March, found technicians who begin by following the aircraft through every station and then specialize down to a single one as volume ramps.  Anduril’s head of production for autonomous airpower explained what that specialization buys.  A technician only has to know about two and a half days’ worth of work.  Thirty workers, trained by the company in California, seeded the entire line.  None of this was decided on the factory floor, because the decisive choices were made years upstream in design, where the company picked aluminum over titanium, a commercial engine over a bespoke one, and mechanical fasteners over specialized bonding, and pushed roughly 90 percent of its components onto commercial supply chains.  A missile fuel vessel gets cast the way large toys are cast.  An airframe is pressed the way fiberglass bathtubs are pressed, which quietly turns every bathtub manufacturer in America into a potential defense supplier.  Notice what happened to training along the way.  It did not disappear so much as move inside the factory and shrink to days, becoming a function of job design rather than a function of the education system.

Four Problems Wearing One Name

Read through a workforce lens, the arsenal now presents at least four distinct problems, and each of them wants a different tool.  The Arsenal floor itself wants almost nothing from the outside world, since job design plus a few days of in-plant instruction covers it, and a two-year degree aimed at that floor answers a question nobody asked.  The supply-chain tier is different.  The casting shops and fabricators the doctrine deliberately recruits will need short, employer-embedded upgrades in quality systems and process discipline, and a community college’s non-credit shop may be exactly the right vendor for that work if it shows up as a vendor instead of a degree program.  The tier Anduril calls exquisite does not go away, and the company concedes as much.  The nuclear attack submarine still needs the ten-year welder, and there the registered apprenticeship and the New England-style pipeline remain precisely the right instruments.  That leaves the mobilization layer, the machinery for converting commercial capacity to defense production in a crisis, which wants a tool that no institution currently owns.  None of this makes community colleges or apprenticeships obsolete.  They are tools, correct in their tiers, and the failure mode belongs to no single institution but to the habit every institution shares, which is pricing its own instrument into every problem it sees.

The Hammer Has an Appropriation

That habit is less a character flaw than an architecture.  Each federal workforce funding stream arrives legally welded to one instrument.  The Workforce Innovation and Opportunity Act (WIOA) buys individual training accounts, Perkins buys college programs, and apprenticeship dollars buy apprenticeships, which means we do not have a workforce system so much as a collection of hammers, each with its own appropriation, and Anduril has just presented a problem that none of them is shaped to hit.  An operator would start from the shape of the demand and then reach into the toolbox.  The funding architecture punishes that instinct, because the money shows up already committed to its tool, and until the dollars can follow the problem, the tools will keep finding nails.

What Anduril’s Doctrine Leaves Unsaid

Anduril’s doctrinal statement invokes Rosie the Riveter, but history has a correction to offer, because Rosie was trained, not found.  The arsenal of democracy ran on Training Within Industry (TWI), a federally coordinated and employer-delivered system that reached 16,511 plants and certified 1,759,650 supervisors between 1940 and 1945, and it did all of that inside the plants rather than through an education system that had no capacity for the job then either.  Anduril has rebuilt the in-plant half of that machine one company at a time and rebuilt it well.  Nobody has rebuilt the coordination half.  Anduril’s own surge story depends on it, since its crisis plan runs through the commercial workforces it names, the bathtub makers and the auto workers, none of whom have been identified, exercised, or organized by anyone.  The work waiting for workforce and economic developers sits exactly there.  Map which regional manufacturers stand one process away from the surge supply chain, let public training dollars follow workers into plants rather than into predetermined programs, and run the conversion exercises that turn a named workforce into a ready one.  Anduril solved the skill problem with design, and the mobilization problem is ours on a clock of the company’s own setting.  Right now, what the Department of War wants is latent surge capacity in the defense manufacturing workforce.  But right now, latent surge capacity is a hypothesis, not a capability, and wars of production are not won on hypotheses.

---